Biofilms form almost everywhere underwater and may take a rusty colour. This is because one mineral excreted by the bacteria is a close relative of rust! It is called ferrihydrite, and is a complicated oxy-hydroxide of oxidised iron.

The bacteria which form this waste product earn their metabolic living by oxidising reduced ferrous iron to oxidised ferric iron, and gaining energy from the process. These bacteria are very widespread in the world. If the pH is very different from a typical lake or river, some other bacteria can even gain energy by the opposite process -  reducing oxidised iron back to reduced iron. Both kinds of bacteria may exist on opposite sides of the same biofilm.

Those scientists who work with ESR (Electron Spin Resonance) are very familiar with its high sensitivity for detection of Fe⁺³ and it immediately occurred to me that it ESR could be a useful tool to examine these iron metabolising bacteria.

The ESR apparatus may not be familiar to readers, so I should say it looks something like two 1 mt diameter electromagnets vertically mounted with a sample space between them, and associated electronics. The sample is put in a thin silica sample tube then into the middle.  A square cross-section tube from the electronics goes into the sample space and bombards the sample with microwaves. The electronics also record if the sample absorbs the microwaves ("resonance"). A lot can be inferred from the precise wavelength this occurs. The technology is generally considered "mature" i.e., well known, and is much used to investigate specific biochemical reactions, as well as in solid-state physics studies.

I am from New Zealand, and have had a long association with the Osaka University ESR laboratory under Prof. Ikeya through a mutual interest in ESR dating, and was going to spend most of 2003 there under a Senior Fellowship offered by the Japanese Society for the Promotion of Science. Biofilms seemed a good project because I could find no previous mention of using ESR to examine them.

There were certainly previous ESR investigations of bacteria in general, and a universal finding was an ESR peak which proved due to oxidation of the lipid membranes of bacteria in air.

It is not possible to take an ESR spectrum of a sample on a metallic film for the kinds of reasons that forbid metal in microwave ovens. The resulting sparking is spectacular and entertaining but bad for the electronics. Therefore it was necessary to grow the biofilms on plastic. Here we went on one of those long diversions which happen quite often in Science, seeming important but not really contributing to the end result. I found that in the literature, the rate of growth of biofilms depended on how polar the film was (how wettable). The more polar or hydrophilic the film, the better the growth. But how does one make a very non-polar surface more polar? (In other words, less like a normal plastic.) It turned out this was already quite well known, because commercial marketers want to print pretty and striking labels and logos on plastic bags, but printers' ink does not stick to plastic well. They found that by carefully controlled exposure to flame the surfaces could be oxidised just enough to hold ink. The process is very common and the result is the brilliant plastic packages which assault our eyes everywhere in supermarkets. We therefore used suitable clear plastic from common packaging.

The joke at the end of this diversion was that we did a comparison with untreated Teflon (which should be very poor for biofilm growth), and found the biofilm grew just as well on the Teflon tape!. This made searching out the specially treated plastic quite pointless.

The ESR spectra from biofilms proved to have an enormous peak due to ferrihydrite, and a small one due to lipid oxidation. The former was obviously from the iron metabolising bacteria and the latter was possibly from all the bacteria present. The iron-metabolising bacteria can be a nuisance in some domestic water supplies, so we were pleased there was potentially some connection of our studies to real problems!

The biofilms contained significant iron, but did not look rusty. This did not really surprise us, because the layers were very thin and hardly visible, but finally made us realise that the iron bacteria were only a small proportion of all the bacteria present, and we came across a paper which mentioned in the Minoh River, our work-site, the majority bacterium was Flavobacteriumjohnsonae, so we were detecting very efficiently a small proportion of the biofilm. The Minoh River flowed through Osaka to the sea, joining larger rivers on the way. It was mainly significant for being a scenic tourist attraction in its upper reaches, and being so accessible, that many colleagues automatically chose it for experiments so there was a moderate sized scientific literature about it.

To our slight embarrassment a short way into the research we discovered a colleague 200 km away in Japan, who knew far more about biofilms than we did, had specialised in the Fe-metabolising biofilms, and was a graduate of Osaka University where we were working. She was rapidly co-opted onto the paper and supplied some fascinating Fe-rich biofilms from various Japanese Hot Springs and cold seeps which we examined using ESR. Some contained hematite (quite a puzzle because normally it is only found in much drier conditions) and one contained magnetite. The magnetic mineral magnetite is increasingly found in biotic tissue, and was already known from some bacterial species.

We left plastic film in the Minoh River to measure biofilm growth with time, and as we feared, suffered various rainstorms which washed the apparatus away, however firmly anchored, forcing a restart several times. The only compensation for these extra field-trips was some superb photos of local dragonflies!

We also analysed biofilms from New Zealand, and found that the results seemed the same as for samples from Japan. Iron-metabolising bacteria exist everywhere, as does iron. Together with bacteria producing manganese dioxide they are responsible for the universal presence of iron and manganese oxides and hydroxides on almost all surfaces underwater, and partly responsible for their negative charge.

The other local problem was the second field site, which was a large stagnant local lake on University property, much used by fishermen. A notice appeared just before the time-series experiments and I correctly inferred from my poor Japanese that forcible removal of all structures in the lake was imminent. The experiments were postponed.

When successful time-series samples had been analysed, we were a little disappointed with the results. Growth as measured by ESR was much more erratic than we would see in a hard sciences experiment, the area where we usually worked.  This showed our bias, because in Biology, the results of experiments are often more scattered than in Physics and this is known as "Biological variability".  We had also expected to see exfoliation. This process is the peeling off of part of a biofilm into a water current as the film becomes mature. But our biofilms appeared to adhere very well to all our plastic films and we saw no exfoliation.

This project was a typical one in which a laboratory has a central machine which many students and staff want to use, and a booking system is necessary. During the individual days on which the machine is available the work is long and intense. The rest of the time is spent extracting the results from charts or computer plots and trying to understand what they may mean, or in field trips.

At first we thought we had found an exceptionally sensitive method, because we could detect film less than a day old, and quite invisible to the eye. However we talked with another expert, Professor Morikawa, in the Tokyo University of Agriculture and Technology who told us that with fluorescent methods 2 hours growth could be detected. This seemed much better than ESR could accomplish, but calculation showed ESR sensitivity was actually very high. So we concluded the method is sensitive for the Fe-bacteria but may not be so sensitive for bacteria in general. With special care we think we might detect a single cell allowed to accumulate iron compounds for a few hours.

In the course of the work, it was found that the sample of Professor Tazaki which contained magnetite, also showed a remarkable angular dependence for the spectrum. This is being written up separately, but seemed to show that in some circumstances ESR could be used rather like X-ray Diffraction, which has intensity peaks at different angles. These give information about the geometrical arrangement of the ESR sites, in this case existing as a kind of pseudocrystal.

The results explain some of what happens when a piece of pollutant plastic is put in a river or stream. More interesting might be the establishment of biofilms on rock, because that might help us to study weathering processes. There is already so much iron in rock we could not use it for ESR experiments, but probably it is possible to find a silicone plastic film which would normally have no iron and could be quite a close imitation of a rock surface.

The biofilms partly explain why iron rusts very rapidly in water and we wondered whether bacteria might be partly responsible for rusting in air. However specimens of rust did not have all the same ESR peaks, particularly the peak due to bacterial membranes, so we concluded that the rusting process in air continues mainly without bacteria.

We were rather surprised not to observe any reduced manganese (Mn⁺²). This gives a strong ESR signal, so we concluded that any present must be the oxidised Mn⁺⁴ form which does not give a strong ESR signal. However this raised a clever possibility for the future. If we had the opportunity to study biofilms in reducing conditions, probably Fe would be reduced and would not give an ESR signal, but Mn would be in the reduced form and easily detected.

In summary the project showed ESR does have real potential for examining biofilms, or at least the parts of biofilms which consist of Fe-metabolising bacteria. If they produce hematite or magnetite besides the usual ferrihydrite, that will be readily detected. The spectra showed that generally the ferrihydrite is amorphous.

The question is whether any group will ever use the results. The timing of the project was at the end of an era. It was the last few months of the ESR career of Professor Ikeya, who retired and in the characteristic Japanese way the people and possessions of his research group were divided amongst other groups in the university. ESR as an independent research subject will not continue.

So we were left asking: Is the end of an era the appropriate time to investigate the potential of a new method?

Note: Electronic Journal of Biotechnology is not responsible if on-line references cited on manuscripts are not available any more after the date of publication.